EEVblog #577 – Precision 1A Current Source Part 2Posted on February 7th, 2014 10 comments
Part 1 HERE
Dave gets his precision 1A current source circuit working on a breadboard.
VPR221Z Precision 4-terminal Z-Foil resistor from Vishay
LTC6655 Voltage Reference
Right at the end of the video you made some measurements for overshoot but I believe there might be a little bit of a gotcha there. The measurement was done on the voltage across the resistor but because of the large cap, that won’t correspond to the load current as the cap is charging. I would suspect there would actually be a large overshoot in current as that capacitor is charging.
May I propose an alternate compensation scheme? Place a resistor in series with a capacitor from the gate of the MOSFET to ground. As a point to start, try a resistor value of 1/10th of the gate resistor.
Would love to see a follow up if you try this, I’m a big fan of anything analog and love these videos, thanks so much!
current flowing into the buffer op-amp will “NOT EVEN HALF A BEE’S DICK” ?!? Did I hear properly ?
We learn not only electronics but colorful linguistics here ! LOL !
My belly laugh of the day !
Poor little MOSFET
newark electronics shows a 10R 1% VPR221ZT as special order, qnty 1, $40 USD!!!
I think I have to second this comment. That’s a very good observation. During power-up, when the cap is discharged, it effectively shorts the current sense resistor, so the sense input of the voltage reference ‘sees’ 0 volts at its sense pin. It will try to compensate for that by increasing the voltage at the drive output all the way until the MOSFET is turned on completely, causing a very significant current spike.
So, as Tim says, I think right at the gate of the MOSFET would be the correct place for a power-up delay circuit. A simple RC-lowpass might even do the trick, even though it will also slow down the load regulation of the circuit.
I was wondering, what would be the effect of putting a small inductance in series with the load? Just like output capacitance stabilizes the output voltage of a voltage regulator, inductance should stabilize the output current of a current regulator. Maybe I am totally off with this idea, so I would love to hear some comments from someone who has more experience with regulator design than I.
Thanks a lot for your videos Dave! Besides the teardowns I really enjoy this type of series of videos, where you show the entire process of troubleshooting a circuit. For me, as a relative beginner in practical electronics, this really helps!
The RFP3055 MOSFET ran hot? Its RDSon is pretty high. Its specs look pretty ordinary. There are many MOSFETs with a much lower RDSon in much small packages.
I like the IRFML8244 for general purpose use when space is a premium. With an IDS of 1A, it will barely get warm with an RDSon of roughly 20 milliohms. It is in a SOT 23 package and is cheap and commonly available. I use fairly thick tracks to the drain and source for higher current applications to keep voltage drops minimal. The tracks also act as a heatsink.
Thanks for the great video. The op-amp did the trick as you say. The app note said 2.2uF would be OK!
It’s not really surprising that the MOSFET ran hot. In this application it does not work as a switch, but as a variable resistor that is controlled by the control loop of the voltage reference. Its purpose is to drop all the voltage that does not drop across the load. So its RDSon does not matter at all.
In this test configuration the load is a ‘short’ (a multimeter’s current input), so there is virtually no voltage drop across the load. This means that with 5V supply voltage and a current of 1A, 1.25V drop across the current sense resistor and 3.75V are left to drop across the MOSFET’s drain-source channel. This gives a power dissipation of about 3.75W. Ouch for the MOSFET and the person touching it
BTW, thanks for the hint towards the IRFML8244. Even here in Germany it is really cheap and its low on-resistance could make it interesting for a couple of applications.
PS: Sorry, my previous comment was intended to be a reply to the first comment to this video.
Don’t forget that the threshold voltage of that mosfet is 2-4V. With Vdd=5V and 1.25V across the shunt in the source, there is only 3.75V left across Vgs, and thats if and only if pin 7 of that chip can go to 5V. It seems to work with that particular transistor, though.
What about leakage current through the 470uF cap? Won’t that increase the overall current as well?
Leave a reply